Abstract 003: Metabolome-Wide Association Study With Habitual Physical Activity Among 4119 U.S. Participants

Background: Being physically active has been associated with lower risks of cardiometabolic diseases. However, the biological mechanisms underlying these associations remain unclear, and can be investigated using a metabolomics approach. Objective: To identify plasma metabolites associated with habitual physical activity in a U.S. population. Methods: Our study population included 4119 participants in the Nurses’ Health Study (NHS), NHS II, and Health Professionals Follow-up Study (HPFS). Physical activity was assessed periodically by self-reported questionnaire starting from 1984 in NHS, 1989 in NHSII, and 1986 in HPFS. We used physical activity measured closest before blood collection as exposure. Metabolic profiling was conducted by liquid chromatography-mass spectrometry (LC-MS). Our study included 287 known metabolites, with 64% of them classified as lipids (58 triglycerides [TAGs], 13 diglycerides [DAGs], 13 cholesteryl esters [CEs], 6 lysophosphatidylethanolamines [LPEs], 37 phosphocholines [PCs], 10 lysophosphatidylcholines [LPCs], 24 phosphatidylethanolamines [PEs], and 25 carnitines). We examined associations of physical activity with plasma metabolites using linear regression, and corrected for multiple testing using tail probability of the proportion of false positives (TPPFP) and accounting for correlated tests using bootstrapping. Results: Using linear model adjusting for age, case-control status, labcode, smoking status, fasting status, alcohol intake, and diet quality, physical activity was significantly associated with 27 metabolites after correcting for multiple testing (TPPFP <0.05). Among the 27 identified metabolites, four CEs, eight PCs, and three LPCs were positively associated with physical activity, and five TAGs and four DAGs were inversely associated. After additionally adjusting for body mass index (BMI), the associations of physical activity with the majority of identified metabolites were attenuated; physical activity remained positively associated with three CEs (C18:2, C16:0, and C18:1), three PCs (C36:4 PC-A, C36:0 PC, C34:3 PC plasmalogen), and one LPC (C18:2). Enrichment analysis showed that physical activity was significantly positively associated with the CE category (P value for enrichment=0.048). Conclusions: We identified seven metabolites that were significantly associated with physical activity after adjusting for BMI, which may help identify biomarkers of physical activity and provide insight into biological mechanisms underlying the beneficial effect of being physically active on cardiometabolic health.